DroughtCast uses ML to ensure food security

Simulated meteorology and satellite dataprovide inputs for a recurrent neural network to accurately forecast up to twelve weeks into the future

Up to 14 percent of cropland around the world has suffered from moderate to extreme levels of drought in 2022. This has had a significant effect on global agricultural supply chains, including decreased crop yields and delayed planting schedules. By 2050, drought could affect 75% of the global population, according to the United Nations (UN). This has potentially alarming consequences for global trade – when water levels are too low to facilitate the movement of vessels, the supply chain breaks down.

India loses 2-5% of GDP to drought

India features on the Global Drought Vulnerability Index, which is part of the assessment. Geographically, India’s drought vulnerability compares with that of sub-Saharan Africa. “The effect of severe droughts was estimated to have reduced India’s gross domestic product by 2-5 per cent over the 20 years from 1998-2017,” the assessment said.

March 2022 was the hottest month in India since the country’s meteorological department started maintaining records 122 years ago. Blistering heat has scorched wheat fields in the country, reducing expected yields by an average of 15 percent and prompting India to ban wheat exports. Europe is emerging from a drought that appears to be the worst in at least 500 years. As a result, relative to the five-year average, the forecasts for EU crop yields for maize, soybean, and sunflowers have been cut by 16 percent, 15 percent, and 12 percent, respectively.

DroughtCast – ML technology to forecast drought

Governments and agriculture companies are focusing on using machine learning technologies to forecast drought and parallelly developing drought-resistant crops. The United States Drought Monitor (USDM), an organisation under the U.S. Department of Agriculture has launched DroughtCast, a machine learning framework, to forecast drought, that operates on the knowledge that recent anomalies in hydrology and meteorology drive future changes in drought conditions.

DroughtCast uses simulated meteorology and satellite observed soil moisture as inputs into a recurrent neural network to accurately forecast the USDM between 1 and 12 weeks into the future. Its analysis shows that precipitation, soil moisture, and temperature are the most important input variables when forecasting future drought conditions.

Satellite tech for early drought warning

The award-winning South Asia Drought Monitoring System (SADMS), developed as a part of the International Water Management Institute’s drought program, comes out with a weekly map of drought conditions, as well as a 15-day drought forecast. The system involves using state-of-the-art satellite technology to develop drought-monitoring and early-warning systems. The systems use models to analyse a range of satellite data on rainfall, snowfall, snow cover, soil moisture, vegetation health and crop yield.

When drought struck South Asia in 2020, two southern states of India used SADMS to develop real-time contingency planning measures (including provision of drought-tolerant seed varieties, rainwater harvesting and spraying of potassium nitrate to alleviate drought stress). The areas in which the measures were enacted were later found to have had significantly higher yields and financial returns than those where they were not.

A digital green revolution

Simultaneously, investments are being made on using data science to develop seeds that grow drought-resistant crops. During the “green revolution” of the 1960s, researchers developed new chemical pesticides and fertilisers along with high-yielding crop varieties that dramatically increased agricultural output. Today as Artificial Intelligence (AI) becomes integrated into agriculture, some crop researchers envisage an agricultural revolution with computer science at the helm.

For instance, a wheat breeder has 200 genetically distinct lines, he or she must decide which lines to breed together to optimise yield, disease resistance, protein content, and other traits. The breeder may know which genes confer which traits, but it’s difficult to decipher which lines to cross in what order to achieve the optimum gene combination. The number of possible combinations is more than the stars in the universe. An operations research approach enables a researcher to solve this puzzle by defining the primary objective and then using optimisation algorithms to predict the quickest path to that objective given the relevant constraints.

Editing genes to create resilient crops

Machine learning algorithms and other AI tools have been used by some companies for about a decade, but they’re becoming more widely available to breeders. Breeders use them to analyse data that reflects what’s known about plant biology – for example, how high or fast a plant grows, and whether it can withstand hot or dry conditions in a region.That data can help AI predict the combination of genes – and the genes that regulate them – that can improve a plant’s tolerance without reducing the yield. AlphaFold, a deep learning system developed by DeepMind that can predict the structure of proteins, was used along with other tools to study the genes and proteins involved in the response of potatoes to elevated temperatures. 

AI can improve the earliest phase of the agricultural lifecycle: creating better crop inputs before seeds are in the ground. For example, gene editing technology CRISPR – another innovation developed in a different industry – could help to design more resilient, high-yield seeds. Companies are applying AI to improve its speed and efficacy. Because many crops are so genetically complex – corn has 32,000 genes compared to 20,000 in humans – AI is invaluable in helping researchers understand the effects of editing multiple genes.

Companies like Inari and Cibus are using these technologies to bump up crop yields while requiring less water and other inputs. Increasing yields of staple crops like corn, soy and wheat is critical as the global population grows and natural disasters like droughts, exacerbated by climate change, make farming more difficult. These innovations could help to stabilise the global food supply when geopolitical events disrupt trade. We’re seeing this play out now.

Know more about the syllabus and placement record of our Top Ranked Data Science Course in Kolkata, Data Science course in Bangalore, Data Science course in Hyderabad, and Data Science course in Chennai.

https://praxis.ac.in/old-backup/data-science-course-in-bangalore/